Semantic segmentation of HeLa cells: An objective comparison between one traditional algorithm and four deep-learning architectures.

The quantitative study of cell morphology is of great importance as the structure and condition of cells and their structures can be related to conditions of health or disease. The first step towards that, is the accurate segmentation of cell structures. In this work, we compare five approaches, one traditional and four deep-learning, for the semantic segmentation of the nuclear envelope of cervical cancer cells commonly known as HeLa cells. Images of a HeLa cancer cell were semantically segmented with one traditional image-processing algorithm and four three deep learning architectures: VGG16, ResNet18, Inception-ResNet-v2, and U-Net. Three hundred slices, each 2000 × 2000 pixels, of a HeLa Cell were acquired with Serial Block Face Scanning Electron Microscopy. The first three deep learning architectures were pre-trained with ImageNet and then fine-tuned with transfer learning. The U-Net architecture was trained from scratch with 36, 000 training images and labels of size 128 × 128. The image-processing algorithm followed a pipeline of several traditional steps like edge detection, dilation and morphological operators. The algorithms were compared by measuring pixel-based segmentation accuracy and Jaccard index against a labelled ground truth. The results indicated a superior performance of the traditional algorithm (Accuracy = 99%, Jaccard = 93%) over the deep learning architectures: VGG16 (93%, 90%), ResNet18 (94%, 88%), Inception-ResNet-v2 (94%, 89%), and U-Net (92%, 56%).

3D profilometry and cell viability studies for drug response screening.

In vitro tests for assessing cell viability and drug response are widely employed for determining cytotoxicity of drugs, chemicals, or material substrates. These assays have some advantages, such as speed, reduced cost, and potential for automation. However, since these tests are often run with a huge amount of cells, the characteristic properties of a single cell can be masked leading to a lack of the diagnostic features of these assays. Vital processes as proliferation and cell death (either necrosis or apoptosis) are associated to drastic changes of volume and surface analysis techniques like 3D optical scanning profilometry allow noninvasive and nondestructive approach with fast detection and good resolution at nano-microscale. Here, we demonstrate how coupling noninvasive morphological surface analysis techniques with well assessed biochemical methods can help to establish the relationship between the modifications on cellular viability induced by precursors of proliferation and cell death and variations on cell volume induced by these treatments. The proposed approach has demonstrated improved efficiency on the assessment of inductive changes on tumoral cells in comparison to non-tumoral cells upon administration of proliferative nontoxic or cytotoxic substances like chemotherapeutics.

A Design of Microfluidic Chip with Quasi-Bessel Beam Waveguide for Scattering Detection of Label-Free Cancer Cells.

Light scattering detection in microfluidic chips provides an important tool to identify cancer cells without any label processes. However, forward small-angle scattering signals of cells, which are related to their sizes and morphologies, are hard to be detected accurately when scattering angle is less than 11° in microfluidic chips by traditional lighting design due to the influence of incident beam. Therefore, cell's size and morphology being the golden standard for clinical detection may lose their efficacy in recognizing cancer cells from healthy ones. In this article, a novel lighting design in microfluidic chips is put forward in which traditional incident Gaussian beam can be modulated into quasi-Bessel beam by a microprism and waveguide. The quasi-Bessel beam's advantages of nondiffraction theoretically make forward scattering (FS) detection less than 11° possibly. Our experimental results for peripheral blood lymphocytes of human beings and cultured HeLa cells show that the detection rates increase by 47.87% and 46.79%, respectively, by the novel designed microfluidic chip compared to traditional Gaussian lighting method in microfluidic chips. © 2019 International Society for Advancement of Cytometry.

Schlieren two-photon microscopy for phase-contrast imaging.

While simultaneous phase-contrast and two-photon fluorescence imaging in microscopy can bring abundant biomedical information, it is difficult to retrieve phase information from conventional two-photon microscopes. To realize low-cost, in situ phase-contrast and two-photon fluorescence imaging, we propose Schlieren two-photon microscopy, a method that implements phase-contrast imaging on two-photon microscopes. This method involves spatially modulated fluorescence plates, which are made of two-photon fluorescence dyes or upconversion nanoparticles. We demonstrate that the fluorescence intensity fluctuation reflects the phase gradients of the specimen via theoretical analysis, simulations, and experiments. The proposed method is fully compatible with commercial two-photon microscopes, thus enabling widespread applications in live tissue imaging.

GPU-based deep convolutional neural network for tomographic phase microscopy with ℓ1 fitting and regularization.

Tomographic phase microscopy (TPM) is a unique imaging modality to measure the three-dimensional refractive index distribution of transparent and semitransparent samples. However, the requirement of the dense sampling in a large range of incident angles restricts its temporal resolution and prevents its application in dynamic scenes. Here, we propose a graphics processing unit-based implementation of a deep convolutional neural network to improve the performance of phase tomography, especially with much fewer incident angles. As a loss function for the regularized TPM, the ℓ1-norm sparsity constraint is introduced for both data-fidelity term and gradient-domain regularizer in the multislice beam propagation model. We compare our method with several state-of-the-art algorithms and obtain at least 14 dB improvement in signal-to-noise ratio. Experimental results on HeLa cells are also shown with different levels of data reduction.

Morphology and dynamics of tumor cell colonies propagating in epidermal growth factor supplemented media.

The epidermal growth factor (EGF) plays a key role in physiological and pathological processes. This work reports on the influence of EGF concentration (c EGF) on the modulation of individual cell phenotype and cell colony kinetics with the aim of perturbing the colony front roughness fluctuations. For this purpose, HeLa cell colonies that remain confluent along the whole expansion process with initial quasi-radial geometry and different initial cell populations, as well as colonies with initial quasi-linear geometry and large cell population, are employed. Cell size and morphology as well as its adhesive characteristics depend on c EGF. Quasi-radial colonies (QRC) expansion kinetics in EGF-containing medium exhibits a complex behavior. Namely, at the first stages of growth, the average QRC radius evolution can be described by a t 1/2 diffusion term coupled with exponential growth kinetics up to a critical time, and afterwards a growth regime approaching constant velocity. The extension of each regime depends on c EGF and colony history. In the presence of EGF, the initial expansion of quasi-linear colonies (QLCs) also exhibits morphological changes at both the cell and the colony levels. In these cases, the cell density at the colony border region becomes smaller than in the absence of EGF and consequently, the extension of the effective rim where cell duplication and motility contribute to the colony expansion increases. QLC front displacement velocity increases with c EGF up to a maximum value in the 2-10 ng ml-1 range. Individual cell velocity is increased by EGF, and an enhancement in both the persistence and the ballistic characteristics of cell trajectories can be distinguished. For an intermediate c EGF, collective cell displacements contribute to the roughening of the colony contours. This global dynamics becomes compatible with the standard Kardar-Parisi-Zhang growth model, although a faster colony roughness saturation in EGF-containing medium than in the control medium is observed.

Gene Therapy for Treatment of Chronic Hyperammonemia in a Rat Model of Hepatic Encephalopathy.

INTRODUCTION AND AIM: Hepatic encephalopathy (HE), caused by hyperammonemia resulting from liver disease, is a spectrum of neuropsychiatric and motor disorders that can lead to death. Existing therapies are deficient and alternative treatments are needed. We have shown that gene therapy with a baculovirus vector containing the glutamine synthetase (Bac-GS) gene is efficient for reducing ammonia levels in an acute hyperammonemia rat model. However, the most common condition resulting from liver disease is chronic hyperammonemia. In this work, Bac-GS was evaluated in bile-duct ligated rats, a chronic liver disease model with hyperammonemia and some characteristics of Type C HE. MATERIAL AND METHODS: Bac-GS was tested for mediating GS overexpression in HeLa cells and H9C2 myotubes. For determining the utility of Bac-GS for the reduction of ammonia levels in a chronic hyperammonemia animal model, four groups of rats were treated: control, sham, ligated with Bac-GS and ligated with Bac-GFP. Baculoviruses were injected i.m. 18 days post-surgery. Blood was drawn 2, 3 and 4 weeks post-surgery and plasma ammonia concentrations were quantified. RESULTS: In protein lysates of cells and myotubes transduced with Bac-GS, a 44 kDa band corresponding to GS was detected. Significant results were obtained in the hyperammonemic bile-duct ligated rat model, as plasma ammonia was reduced to normal levels 3 days after treatment with Bac-GS. Furthermore, a transitory effect of Bac-GS was observed. CONCLUSION: Our results show that gene therapy by delivering GS is a promising alternative for treatment of hyperammonemia in acute-on-chronic liver failure patients with HE.

Enhancing image contrast of carbon nanotubes on cellular background using helium ion microscope by varying helium ion fluence.

Carbon nanotubes (CNTs) have become an important nano entity for biomedical applications. Conventional methods of their imaging, often cannot be applied in biological samples due to an inadequate spatial resolution or poor contrast between the CNTs and the biological sample. Here we report a unique and effective detection method, which uses differences in conductivities of carbon nanotubes and HeLa cells. The technique involves the use of a helium ion microscope to image the sample with the surface charging artefacts created by the He+ and neutralised by electron flood gun. This enables us to obtain a few nanometre resolution images of CNTs in HeLa Cells with high contrast, which was achieved by tailoring the He+ fluence. Charging artefacts can be efficiently removed for conductive CNTs by a low amount of electrons, the fluence of which is not adequate to discharge the cell surface, resulting in high image contrast. Thus, this technique enables rapid detection of any conducting nano structures on insulating cellular background even in large fields of view and fine spatial resolution. The technique demonstrated has wider applications for researchers seeking enhanced contrast and high-resolution imaging of any conducting entity in a biological matrix - a commonly encountered issue of importance in drug delivery, tissue engineering and toxicological studies.

Cell Detection Using Extremal Regions in a Semisupervised Learning Framework.

This paper discusses an algorithm to build a semisupervised learning framework for detecting cells. The cell candidates are represented as extremal regions drawn from a hierarchical image representation. Training a classifier for cell detection using supervised approaches relies on a large amount of training data, which requires a lot of effort and time. We propose a semisupervised approach to reduce this burden. The set of extremal regions is generated using a maximally stable extremal region (MSER) detector. A subset of nonoverlapping regions with high similarity to the cells of interest is selected. Using the tree built from the MSER detector, we develop a novel differentiable unsupervised loss term that enforces the nonoverlapping constraint with the learned function. Our algorithm requires very few examples of cells with simple dot annotations for training. The supervised and unsupervised losses are embedded in a Bayesian framework for probabilistic learning.

In vitro investigation of clofazimine analogues for antiplasmodial, cytotoxic and pro-oxidative activities.

BACKGROUND: Tetramethyl-piperidine-substituted, B4119 and B4158 have been shown to exhibit antiplasmodial activity. OBJECTIVES: The in vitro antiplasmodial, cytotoxic and oxidative activities of clofazimine and its analogues, all TMP (tetramethylpiperidyl)-substituted phenazines except B669, were evaluated in this study. METHODS: The antiplasmodial activity of the compounds against RB-1 and pfUP10 laboratory strains of Plasmodium falciparum was investigated by flow cytometry. The cytotoxic activity against HeLa cells and oxidative activity were studied employing colorimetric and cytochrome C reduction assays respectively. RESULTS: The riminophenazine agents exhibited antiplasmodial action of varying degrees: B669, B4100 and B4103 showed the best activity while B4121 and B4169 exhibited significant activity at 2µg/ml. Clofazimine had no antiplasmodial activity. The compounds B4100, B4103, B4121 and B4169 exhibited significant cytotoxic activity against HeLa cells at concentrations of 0.5µg/ml and above while B669 was active at 2µg/ml. Clofazimine and B669 tested at a concentration of 0.5µg/ml caused enhancement (p ≤ 0.05) of neutrophil superoxide production when compared to the FMLP control while all the other TMP-derivatives had no effect (p ≥ 0.05). CONCLUSION: Tetramethylpiperidyl-subsituted phenazines may potentially be useful antimalarial/antitumor agents with no pro-oxidative properties. In vivo studies on the agents relative to these properties are recommended.

Size- and speed-dependent mechanical behavior in living mammalian cytoplasm.

Active transport in the cytoplasm plays critical roles in living cell physiology. However, the mechanical resistance that intracellular compartments experience, which is governed by the cytoplasmic material property, remains elusive, especially its dependence on size and speed. Here we use optical tweezers to drag a bead in the cytoplasm and directly probe the mechanical resistance with varying size a and speed V We introduce a method, combining the direct measurement and a simple scaling analysis, to reveal different origins of the size- and speed-dependent resistance in living mammalian cytoplasm. We show that the cytoplasm exhibits size-independent viscoelasticity as long as the effective strain rate V/a is maintained in a relatively low range (0.1 s-1 < V/a < 2 s-1) and exhibits size-dependent poroelasticity at a high effective strain rate regime (5 s-1 < V/a < 80 s-1). Moreover, the cytoplasmic modulus is found to be positively correlated with only V/a in the viscoelastic regime but also increases with the bead size at a constant V/a in the poroelastic regime. Based on our measurements, we obtain a full-scale state diagram of the living mammalian cytoplasm, which shows that the cytoplasm changes from a viscous fluid to an elastic solid, as well as from compressible material to incompressible material, with increases in the values of two dimensionless parameters, respectively. This state diagram is useful to understand the underlying mechanical nature of the cytoplasm in a variety of cellular processes over a broad range of speed and size scales.

Comparative Study on Two Different Methods for Determination of Hydraulic Conductivity of HeLa Cells During Freezing.

BACKGROUND: The measurement of hydraulic conductivity of the cell membrane is very important for optimizing the protocol of cryopreservation and cryosurgery. There are two different methods using differential scanning calorimetry (DSC) to measure the freezing response of cells and tissues. Devireddy et al. presented the slow-fast-slow (SFS) cooling method, in which the difference of the heat release during the freezing process between the osmotically active and inactive cells is used to obtain the cell membrane hydraulic conductivity and activation energy. Luo et al. simplified the procedure and introduced the single-slow (SS) cooling protocol, which requires only one cooling process although different cytocrits are required for the determination of the membrane transport properties. To the best of our knowledge, there is still a lack of comparison of experimental processes and requirements for experimental conditions between these two methods. This study made a systematic comparison between these two methods from the aforementioned aspects in detail. METHODS: The SFS and SS cooling methods mentioned earlier were utilized to obtain the reference hydraulic conductivity (Lpg) and activation energy (ELp) of HeLa cells by fitting the model to DSC data. RESULTS: With the SFS method, it was determined that Lpg = 0.10 µm/(min·atm) and ELp = 22.9 kcal/mol; whereas the results obtained by the SS cooling method showed that Lpg = 0.10 µm/(min·atm) and ELp = 23.6 kcal/mol. CONCLUSIONS: The results indicated that the values of the water transport parameters measured by two methods were comparable. In other words, the two parameters can be obtained by comparing the heat releases between two slow cooling processes of the same sample according to the SFS method. However, the SS method required analyzing heat releases of samples with different cytocrits. Thus, more experimental time was required.

Divergent microtubule assembly rates after short- versus long-term loss of end-modulating kinesins.

Depletion of microtubule (MT) regulators can initiate stable alterations in MT assembly rates that affect chromosome instability and mitotic spindle function, but the manner by which cellular MT assembly rates can stably increase or decrease is not understood. To investigate this phenomenon, we measured the response of microtubule assembly to both rapid and long-term loss of MT regulators MCAK/Kif2C and Kif18A. Depletion of MCAK/Kif2C by siRNA stably decreases MT assembly rates in mitotic spindles, whereas depletion of Kif18A stably increases rates of assembly. Surprisingly, this is not phenocopied by rapid rapamycin-dependent relocalization of MCAK/Kif2C and Kif18A to the plasma membrane. Instead, this treatment yields opposite affects on MT assembly. Rapidly increased MT assembly rates are balanced by a decrease in nucleated microtubules, whereas nucleation appears to be maximal and limiting for decreased MT assembly rates and also for long-term treatments. We measured amplified tubulin synthesis during long-term depletion of MT regulators and hypothesize that this is the basis for different phenotypes arising from long-term versus rapid depletion of MT regulators.

Direct and sustained intracellular delivery of exogenous molecules using acoustic-transfection with high frequency ultrasound.

Controlling cell functions for research and therapeutic purposes may open new strategies for the treatment of many diseases. An efficient and safe introduction of membrane impermeable molecules into target cells will provide versatile means to modulate cell fate. We introduce a new transfection technique that utilizes high frequency ultrasound without any contrast agents such as microbubbles, bringing a single-cell level targeting and size-dependent intracellular delivery of macromolecules. The transfection apparatus consists of an ultrasonic transducer with the center frequency of over 150 MHz and an epi-fluorescence microscope, entitled acoustic-transfection system. Acoustic pulses, emitted from an ultrasonic transducer, perturb the lipid bilayer of the cell membrane of a targeted single-cell to induce intracellular delivery of exogenous molecules. Simultaneous live cell imaging using HeLa cells to investigate the intracellular concentration of Ca(2+) and propidium iodide (PI) and the delivery of 3 kDa dextran labeled with Alexa 488 were demonstrated. Cytosolic delivery of 3 kDa dextran induced via acoustic-transfection was manifested by diffused fluorescence throughout whole cells. Short-term (6 hr) cell viability test and long-term (40 hr) cell tracking confirmed that the proposed approach has low cell cytotoxicity.

Photophysics of a Live-Cell-Marker, Red Silicon-Substituted Xanthene Dye.

Dyes with near-red emission are of great interest because of their undoubted advantages for use as probes in living cells. In-depth knowledge of their photophysics is essential for employment of such dyes. In this article, the photophysical behavior of a new silicon-substituted xanthene, 7-hydroxy-5,5-dimethyl-10-(o-tolyl)dibenzo[b,e]silin-3(5H)-one (2-Me TM), was explored by means absorption, steady-state, and time-resolved fluorescence. First, the near-neutral pH, ground-state acidity constant of the dye, pKN-A, was determined by absorbance and steady-state fluorescence at very low buffer concentrations. Next, we determined whether the addition of phosphate buffer promoted the excited-state proton-transfer (ESPT) reaction among the neutral and anion form of 2-Me TM in aqueous solutions at near-neutral pH. For this analysis, both the steady-state fluorescence method and time-resolved emission spectroscopy (TRES) were employed. The TRES experiments demonstrated a remarkably favored conversion of the neutral form to the anion form. Then, the values of the excited-state rate constants were determined by global analysis of the fluorescence decay traces recorded as a function of pH, and buffer concentration. The revealed kinetic parameters were consistent with the TRES results, exhibiting a higher rate constant for deprotonation than for protonation, which implies an unusual low value of the excited-state acidity constant pK*N-A and therefore an enhanced photoacid behavior of the neutral form. Finally, we determined whether 2-Me TM could be used as a sensor inside live cells by measuring the intensity profile of the probe in different cellular compartments of HeLa 229 cells.

Effects of temperature and cellular interactions on the mechanics and morphology of human cancer cells investigated by atomic force microscopy.

Cell mechanics plays an important role in cellular physiological activities. Recent studies have shown that cellular mechanical properties are novel biomarkers for indicating the cell states. In this article, temperature-controllable atomic force microscopy (AFM) was applied to quantitatively investigate the effects of temperature and cellular interactions on the mechanics and morphology of human cancer cells. First, AFM indenting experiments were performed on six types of human cells to investigate the changes of cellular Young's modulus at different temperatures and the results showed that the mechanical responses to the changes of temperature were variable for different types of cancer cells. Second, AFM imaging experiments were performed to observe the morphological changes in living cells at different temperatures and the results showed the significant changes of cell morphology caused by the alterations of temperature. Finally, by co-culturing human cancer cells with human immune cells, the mechanical and morphological changes in cancer cells were investigated. The results showed that the co-culture of cancer cells and immune cells could cause the distinct mechanical changes in cancer cells, but no significant morphological differences were observed. The experimental results improved our understanding of the effects of temperature and cellular interactions on the mechanics and morphology of cancer cells.

Fourier Transform Infrared spectroscopy discloses different types of cell death in flow cytometrically sorted cells.

Fourier Transform Infrared (FTIR) spectroscopy is a label free methodology showing promise in characterizing different types of cell death. Cervical adenocarcinoma (HeLa) and African monkey kidney (Vero) cells were treated with a necrosis inducer (methanol), novel apoptotic inducers (diphenylphosphino gold (I) complexes) and positive control, auranofin. Following treatment, cells stained with annexin-V and propidium iodide were sorted using a Fluorescence Activated Cell Sorter (FACS Aria) to obtain populations consisting of either viable, necrotic or apoptotic cells. Transmission Electron Microscopy confirmed successful sorting of all three populations. Four bands were identified which could discriminate between viable and necrotic cells namely 989 cm(-1), 2852 cm(-1), 2875 cm(-1) and 2923 cm(-1). In HeLa cells viable and induced apoptosis could be distinguished by 1294 cm(-1), while four bands were different in Vero cells namely; 1626 cm(-1), 1741 cm(-1), 2852 cm(-1) 2923 cm(-1). Principal Component Analysis showed separation between the different types of cell death and the loadings plots indicated an increase in an additional band at 1623 cm(-1) in dead cells. FTIR spectroscopy can be developed into an invaluable tool for the assessment of specific types of chemically induced cell death with notably different molecular signatures depending on whether the cells are cancerous and mechanism of cell death.

Verifying the function and localization of genetically encoded Ca2+ sensors and converting FRET ratios to Ca2+ concentrations.

Genetically encoded, ratiometric, fluorescent Ca(2+) biosensors can be used in living cells to quantitatively measure free Ca(2+) concentrations in the cytosol or in organelles. This protocol describes how to perform a calibration of a Ca(2+) sensor expressed in cultured mammalian cells as images are acquired using a widefield fluorescence microscope. This protocol also explains how to calculate Förster resonance energy transfer (FRET) ratios from acquired images and how to convert FRET ratios to Ca(2+) concentrations.

Surface plasmon coupled emission in micrometer-scale cells: a leap from interface to bulk targets.

Surface plasmon coupled emission (SPCE) technique has attracted increasing attention in biomolecular interaction analysis and cell imaging because of its high sensitivity, low detection volume and low fluorescence background. Typically, the working range of SPCE is limited at nanometers to an interface. For micrometer-scale samples, new SPCE properties are expected because of complex coupling modes. In this work, cells with different subregions labeled were studied using a SPCE spectroscopy system. Angular and p-polarized emission was observed for cell membrane, cytoplasm, and nucleus labeled with DiI, Nile Red, and propidium iodide, respectively. The SPCE signals were always partially p-polarized, and the maximum emission angle did not shift, regardless of variations in emission wavelength, fluorophore distribution and stained layer thickness. Additionally, increased polarization and a broader angle distribution were also observed with an increase in sample thickness. We also investigated the impact of metallic substrates on the SPCE properties of cells. Compared with Au and Ni substrates, Al substrates presented better polarization and angle distribution. Moreover, the real-time detection of the cell labeling process was achieved by monitoring SPCE intensity. These findings expand SPCE from a surface technique to a 3D method for investigating bulk targets beyond the nanoscale interfaces, providing a basis to apply this technique to study cell membrane fluidity and biomolecule interactions inside the cell and to distinguish between cell subregions.